Extract images from PDF to Markdown
Diagrams, screenshots and figures are part of the document. When you convert a PDF to Markdown here, images come along: embedded inline so they render anywhere, or kept as light placeholders when you only need the text.
Figures travel with the text
A PDF mixes text and images on every page. The converter pulls both: the prose becomes clean Markdown, and each figure is carried over as an image. You choose how, with an image mode. Embed writes images inline (for example as base64) so a single Markdown file renders complete in any viewer. Placeholder keeps the file small and just marks where each image sits, ideal when the text is what you need. Any words inside an image, like a labelled diagram, are picked up by OCR.
Convert an image-heavy PDF
No account needed. Pick the image mode that suits where the Markdown is going.
Open the converter
Install the Chrome extension or open the web app.
Pick embed or placeholder
In settings, choose whether images are embedded inline or kept as placeholders. Embed for a self-contained file, placeholder for the lightest text.
Add the PDF and wait
Drag in the file or paste a URL. Status goes queued, processing, ready. Figures are carried over and on-page text is OCR'd.
Copy or download
Preview the rendered Markdown with images, then copy it or download a .md file.
Embed or placeholder
The right choice depends on whether you want a complete document or just the words.
Embed inline
Placeholders
What each mode produces
Both modes use the same document structure; they differ only in how each figure is written.
With embed, a figure is written inline as a standard Markdown image, so a single .md file is fully self-contained and renders the picture anywhere:
## Figure 2. System architecture  The request flows from the client through the API facade to the worker.
With placeholder, the same spot gets a lightweight marker instead of the image bytes, keeping the file small and easy to diff while still recording where the figure belongs:
## Figure 2. System architecture [image: Figure 2. System architecture] The request flows from the client through the API facade to the worker.
A note on size: embedding an image inline (as base64) can grow the Markdown file several times over for an image-heavy PDF, which is great for a portable document but heavy for an LLM prompt. If you are feeding the text to a model or a search index, placeholders keep the token count down while preserving the captions and headings that carry the meaning.
Common questions
Does it keep images and figures from a PDF?
Yes. Figures and images are carried into the Markdown, either embedded inline (for example as base64) or as a placeholder reference, depending on the image mode you choose.
How are images embedded in the Markdown?
Inline images use standard Markdown image syntax, so they render in any Markdown viewer. A placeholder instead keeps the file lighter and marks where the image belongs.
Can I turn image embedding off?
Yes. An image-mode setting lets you embed images or use placeholders, so you can keep the Markdown small when you only need the text.
What about a fully scanned, image-only page?
Text on the page is OCR'd into Markdown. Where a page is essentially an image with no recoverable text, it can be kept as a page image so nothing is lost. See scanned PDF to Markdown.
Are diagrams and charts converted to data?
Images stay as images; the surrounding captions, labels and any readable text are converted to Markdown. A chart is not turned back into its underlying numbers.
Is it free?
Yes. Convert anonymously in the browser on the free tier (3 slots, 10 MB files, a 15-minute time budget, 1-hour retention). Paid tiers raise every limit.